Agent Memory Is the Next Great Moat

"We're AI-First!" No You're Not. Here's the Test.

A CEO told Justin Watt his company was ready for AI. "We've got our data architecture together," he said, giddy. Justin asked to see it. The guy pulled up an Excel file. The filename? Data Lake. That moment is the whole episode in miniature. Justin Watt, co-founder of Switchboard, studied psychology, not computer science, and that turns out to be his unfair advantage. After stints at IBM and MetaLab (where his teams built products for Uber and Amazon and helped design Slack), Justin realized the hardest part of every technology project is never the technology. It's the humans. Every business challenge is a human challenge wearing a software costume. Switchboard works with mid-market companies, the $50 million to $500 million crowd, the businesses old enough to have 40 years of legacy process and young enough to actually change. These companies think they're AI-enabled because they bought everyone a Claude license. Meanwhile, month-end close runs through one person's spreadsheet that nobody else can read, and if that person quits, the business forgets how it works. Justin's fix is unglamorous and devastatingly effective: map the real workflow, not the org-chart version. Find where humans are doing machine work. Inject AI at the steps where it actually moves the needle. Keep humans in the loop everywhere else. The result isn't layoffs, it's smart people finally doing smart work. In Justin's experience, less than 5% of leadership conversations are about cutting headcount. The conversation is always about the endless pile of work standing between the company and its goals. Along the way, Ryan and Justin cover the AI washing epidemic (blaming layoffs on AI to cover up old hiring mistakes), why frontier lab doom marketing blew up in everyone's faces, the death of "bring your whole self to work," quiet quitting as cowardice, Garth Brooks selling his catalog for a rumored $2 billion, ravens that speak English, and the most surreal government website in existence. * https://withswitchboard.com [https://withswitchboard.com] * https://www.linkedin.com/in/wattjustin/ [https://www.linkedin.com/in/wattjustin/] * https://aiforfounders.co [https://aiforfounders.co] * https://inboxalchemy.co [https://inboxalchemy.co] * https://spcai.org [https://spcai.org] * https://www.war.gov/ufo [https://www.war.gov/ufo] (referenced as war.gov/ufo) * https://suno.com [https://suno.com] (Suno, the AI music generator discussed)

Agent Memory Is the Next Great Moat

What if the dumbest thing your startup does this year is hire? In Zurich, a six-person company is serving Fortune 500 clients with a rule that sounds like heresy: no human in the company can be assigned a task. The software literally locks them out. Every task goes to an agent first, and the agent decides when a human's judgment is actually worth the interruption. That company is Salfati Group, and its founder is Elon Salfati. Yes, Elon. No, not that one. This Elon is a former Israeli intelligence engineer, ex R&D Director at web security firm Reblaze, co-founder of RELE.AI, founder of intelligent testing startup Metiss, and now a PhD researcher in AI security. He has spent his career deleting more code than he writes, and now he is deleting org charts. The episode opens with a ripped-from-the-headlines jump off: Microsoft's Build 2026 announcement of Autopilots, always-on agents with their own identity that act on your behalf. Ryan asks the uncomfortable question: if 10,000 enterprises flip on the same agents, does diversity of thought dissolve into a hive mind? Elon's answer reframes the whole AI transformation conversation. Most companies are stuck sprinkling AI to please the board or deploying point solutions on annoying spreadsheets. The real unlock is flipping the entire model from "a human with an army of agents" to "an army of agents with a human." From there the conversation gets practical, then philosophical, then back again. Elon walks through a real client engagement: a service marketplace with a 51-step quote-to-cash process bleeding retention, and how color coding every step revealed exactly where humans add value and where they were just hands on keyboard. Then Ryan, a lifelong meditator and self-described student of human consciousness, pulls Elon into the deep end: what does it mean that Salfati Group calls its agents sentient? Elon's answer centers on memory, causality, and temporal understanding, and why he believes agent memory is the next great moat. Plants, cats, the Library of Alexandria, and Mr. Bridgewater the Denver farrier all make appearances. It is that kind of episode. * salfati.group * aiforfounders.co * inboxalchemy.co * Elon Salfati on LinkedIn: linkedin.com/in/elonsalfati * Ryan Estes on LinkedIn: linkedin.com/in/estesryan

What it do!? The Jujitsu Secret That Scales Companies Without Force

Two purple belts walk back onto the mat after years away, and the guy who is slow, mindful, and refuses to break a sweat starts sweeping and submitting the meatheads who are gassing out around him. That is not a jujitsu story. That is the whole episode. This week Ryan Estes and Jason Katz, co-founder of Kindling Solutions, skip the warmup and go straight into the thing every founder feels but rarely says out loud: the ground is moving under all of us, the tools are getting absurdly good, and the people winning are not the fastest or the strongest. They are the ones with stillness, leverage, and an authentic voice that no model can fake. Jason walks through the operating system he is quietly building around himself. A morning brief that reads his Slack, his Teams, yesterday's calls, today's calendar, and his open tasks, then hands him a five-line executive summary before he has even left the porch. A content pipeline that researches ideas, scores them, then interviews him in his own voice like Joe Rogan would, so the output is actually him and not another pile of generated mush. Then the conversation turns to the uncomfortable truth he calls his thorn: everyone is an AI consultant now, the way everyone in Colorado had a grow in 2009, and the single-workflow microservices people are productizing today will cost three dollars on a phone very soon. The question is not whether you can automate something. The question is where your margin lives once the press-a-button version arrives. Ryan counters with the optimist's case. The bigger the frontier models get, the wider the gap between AI-native founders and everyone else, and the more demand there is for people who can actually integrate this stuff with taste. His own proof: web traffic, newsletters, and podcasting, the three compounding channels he believes are the only distribution worth grinding for, because you own them and no single algorithm can switch them off overnight. It is fast, funny, and genuinely useful, with a side of hair powder and an au pair from South Africa. Pull up a chair. * https://kindlingsolutions.com [https://kindlingsolutions.com] (relaunching, was not live at recording) * https://www.linkedin.com/in/jasonkatz99/ [https://www.linkedin.com/in/jasonkatz99/] * https://www.headset.io [https://www.headset.io] * https://www.anthropic.com [https://www.anthropic.com] * https://search.google.com/search-console [https://search.google.com/search-console] * https://www.eastonbjj.com [https://www.eastonbjj.com] * https://trynina.co [https://trynina.co] * https://aiforfounders.co [https://aiforfounders.co] * https://inboxalchemy.co [https://inboxalchemy.co] * https://www.linkedin.com/in/estesryan/ [https://www.linkedin.com/in/estesryan/] * https://trynina.co [https://trynina.co] * https://ainativestudent.com [https://ainativestudent.com]

The $18 Billion Backdoor: How One Aussie Founder Plans to Eat LinkedIn Alive

The inbox is dead, and the people who keep emailing it the hardest are the ones killing it fastest. David Connors has watched this happen up close. He sold his recruiting automation startup, Automately, to Sequoia Capital, spent two years inside the firm building tools so its investors and founders could answer one maddening question, "who do we actually know at company X," and walked out with a conviction that turned into a company. That company is The Swarm, the relationship intelligence platform he now runs as Co-Founder and CEO, and this is his second time on the show. In the twelve months since his last visit, the world sped up, the spam cannons got louder, and David got quieter, more grounded, and 35 pounds lighter. The throughline of this episode is simple and a little uncomfortable: AI made it trivially easy to write the perfect cold message, which means the perfect cold message is now worth almost nothing. What it cannot fake is trust. And trust, David argues, is the only currency left. The conversation moves from the personal to the tactical and back again. David opens up about how he protects his attention as a father of two with a third on the way, why he treats work like a sprinter treats a race rather than a marathoner who never stops, and why running yourself into the ground produces expensive decisions you pay for twice. Then Ryan steers into the meat: how The Swarm passively maps the network sitting around your entire company, not just your personal Rolodex, and turns it into a third sales channel that is neither inbound nor outbound. The numbers do the talking. A warm intro converts ten to twenty times better than cold. Google and Microsoft are now filtering out senders you do not recognize. The motion that used to eat ten to 15 hours a week of someone's time now takes ten minutes with agents. And the whole thing compounds, because every customer you close maps a new network you can map next. There is a bigger swing underneath all of it. David is not trying to be a $10 million enrichment-data business. He wants to carve into LinkedIn's roughly $18 billion revenue run rate by building the relationship graph that agents can actually use, the thing LinkedIn built for the SaaS era but will never open up. Whether you buy the vision or not, the practical takeaway lands either way: map your network, treat it like an asset, batch your asks, close the loop, and never become the neighbor who only knocks when they need an egg. https://www.theswarm.com [https://www.theswarm.com] https://www.linkedin.com/in/connorsdavid/ ⁠⁠https://www.linkedin.com/in/estesryan/⁠⁠ ⁠⁠https://aiforfounders.co⁠⁠ https://inboxalchemy.co/ https://trynina.co/ https://ainativestudent.com/ https://www.ymcasf.org [https://www.ymcasf.org]

Stop Getting Paid for Hours. Start Getting Paid for Outcomes.

She didn't get maternity leave. So she rewired the entire way she worked, and accidentally redrew the map of what her career was worth. In 2020, Ashley Gross was just another marketer pulling 80 hours inside a 40-hour job, operating on the oldest visibility hack in the book: be the first one in, be the last one out. Then she became a mom with no leave on the table, and the math stopped working. So she started teaching AI to do her busywork, not to impress anyone, but to claw back time with her newborn. What happened next is the real story. The automation didn't free her. It exposed her. All the work she clawed back came flooding right back in, because she was the comfort person, the human Google, the one who knew where every document lived. Knowledgeable, indispensable, and quietly underpaid. That gap, between the work you do and the work people can see, became her whole thesis. She walked into the CMO's office, handed over her playbook, and built her own unpaid internal AI champion role to test one question: does this knowledge transfer to other humans? Within three months, the answer was a $25 million pipeline overachievement. That was the moment the imposter syndrome died. Then came the newsletter, zero to over 5,000 in two months of cringey, daily, ego-at-the-door posting. Then a Maven waitlist of more than a thousand people telling her they would pay. Only then did she jump, never risking the paycheck that fed her family until the runway was already built. Today AI Workforce Alliance runs on a team of twelve full-timers, ten-plus part-timers, freelancers, and a few agents quietly handling the admin in between. Notion is the centralized brain. Claude and MCP connectors do the talking. The tech stack went from sprawling to five tools. The plan for 2026 is to 10X through partnerships. And she still hates social media, which is exactly why you should trust her when she says you have to do it anyway. * https://aiworkforcealliance.com [https://aiworkforcealliance.com] * https://www.linkedin.com/in/theashleygross [https://www.linkedin.com/in/theashleygross] * The AI Work Week (Wiley), pre-order on Amazon and https://www.barnesandnoble.com [https://www.barnesandnoble.com] * https://tgpdenver.org [https://tgpdenver.org] (The Gathering Place, Denver) * ⁠⁠https://www.linkedin.com/in/estesryan/⁠⁠ * ⁠⁠https://aiforfounders.co⁠⁠ * https://inboxalchemy.co/ * https://trynina.co/ * https://ainativestudent.com/